inline image

In this occasional series we record the views and personal experience of people who have specially contributed to the evolution of ideas in the Journal’s field of interest.

Addiction (A): Walt, what is your professional background and how did you become involved with alcohol studies and the Alcohol Research Group (ARG) at Berkeley?

Walter Clark (WC): My background and studies in graduate school at Berkeley had nothing to do with alcohol. When I entered the alcohol field I was studying under Irving Goffman, conducting observational research on people’s behavior. I had begun my doctoral dissertation, and had passed my qualifying examinations. I was then hired by Ira Cisin as a research assistant to work in the California Drinking Practices Study that he and Genevieve Knupfer were conducting in Berkeley. This was supposed to be a summer job, but as you know, lasted much longer than that. After working with Ira and Genevieve for a while I left to work in Canada. I was offered a job as an assistant professor in a Canadian university and went there with the notion of finishing my dissertation and pursuing an academic career, which I did off and on for about 7 years. I never totally cut my ties with Ira and Genevieve because I came back to Berkeley to work with them again during this period. Still, at some time during that period I decided I did not want an academic career and I had no intention of finishing my doctoral studies. I decided that I would become a researcher, more in the ‘epidemiological mode’ of description, and with much less emphasis on social sciences.


  1. Top of page
  5. References

A: What were you hired to do in this job with Genevieve and Ira?

WC: Ira was a statistical researcher, and Genevieve Knupfer was a sociologist and a psychiatrist. I had never thought about alcohol before other than as a consumer, and neither had these other people, by the way. Neither Genevieve nor Ira had had any previous interest in the study of alcohol, and in fact there were not many such studies in those days. One of my first tasks was to collect and review everything that had been carried out in research on alcohol use in general populations. There had been a study by Riley & Marden [1]. The Gallup Polls had been asking some questions on alcohol use in the 1930s. In fact, they did so until well after the Second World War. They asked a general question worded more or less as ‘do you have occasion to use alcohol beverages or are you a complete abstainer?’. Anyway, in 1947 and 1948 Riley, Marden & Lifshitz [2] asked about frequency of drinking but not about quantity. In 1952 Maxwell [3], in Washington State, studied a sample of the general population. Straus & Bacon [4] were studying college students, and asked for the first time about alcohol problems. Mulford & Miller [5] and colleagues in Iowa also studied general population samples in about 1963 or 1964. That was about all there was to be found, with the exception of studies that had been performed with clinical samples and, of course, most important among those was Jellinek’s study of members of Alcoholics Anonymous (AA).

A: How did Ira and Genevieve share responsibilities?

WC: Genevieve had much more to do with the content of the questionnaires and what we should be looking at. Ira was concerned with drawing an area probability sample of the city of Berkeley. This task involved me going out and walking around blocks that had been selected randomly and noting every ‘nth’ house and writing the address down on a piece of paper. Later, I returned and knocked on the doors and obtained, where I could, the names of all the people living there and randomly selected one, who was to be interviewed. And then I became the interviewer. All these were face-to-face interviews, using either close-ended questions or recording verbatim what people had to say. Later we would code all that information for analysis. The sample was several hundreds large, but as time went by we had more interviewers and more research assistants.

A: How did you handle the data analysis?

WC: Once we had all the data in hand, one thing we could not do was key-punch all the IBM or Hollerith cards. None of us were good typists much less good key-punch operators, so we sent that out; but everything else was done by hand, and when the cards came back the data were analyzed by running the cards through a counter-sorter like decks of cards. We would then count the results that the sorter showed, and with a slide-rule and mechanical calculator write pages and pages and pages of percentage calculations of this. We occasionally did try more advanced analytical techniques such as factor analysis It was a time-consuming effort. Eventually you reach the theological stage of trying to name what you think the factors mean. This would take 3 or 4 days and at the end of the calculations, what was uppermost in mind was the hope that all the mistakes that were most certainly made in these calculations did not distort the results too much. But then again, that should not bother you overly much because you knew well that no one was going to do the same thing over again to make sure that your calculations were correct.

A: Response rates?

WC: The interviews were easy to obtain and people were more than willing to cooperate. We thought a 90% response rate was just acceptable, and it was possible to obtain 95% or 96% with a good number of calls back to obtain people’s agreement and catch them at home.

A: What was the main purpose of the California Drinking Practice Study

WC: I do not think that was very clear in anyone’s mind. It went something like this: we knew that alcoholics, whoever they were, were people who drank a great deal. They drank a great deal more than ordinary people, but no one knew what the ordinary people were drinking. The hope was that we would find a normal distribution of alcohol use in a population and then something very different that would describe this population of people who were going to end up in clinics or dying or cirrhosis. People had a vague idea that they were searching for something which would somehow separate normal drinkers from those who were destined to have serious problems. It was all very exploratory.

A: So the hope was that there was a bimodal distribution, with alcoholics grouped at the right end of the distribution?

WC: I do not know that anyone put it into these words, but something like that was probably underlying the notion of those who funded the study, and perhaps it was also in the minds of Ira and Genevieve.


  1. Top of page
  5. References

A: What types of studies followed the California Drinking Practices Study at the ARG?

WC: Much of my knowledge applies to the 1970s, the time after I came back from Canada to the United States. At that time the people at ARG which, by the way, was then called the Social Research Group, were Don Cahalan, Robin Room and some others. They encouraged me to come back on a full-time basis, which I did until I retired in the mid-1980s. Don had had an affiliation with Ira Cisin at the George Washington University. Ira then returned to that university to resume his career there, and asked Don to become the director of the alcohol studies being conducted at the time. This was now a combination of national studies and the ongoing studies of the California group. I hired Robin to work as a research assistant, under similar circumstances to mine, for a summer job. Robin came to alcohol research after being discouraged with the study of English literature. When we met, Robin Room asked me for a job and being an honest man said ‘but I know nothing about sociology’. I replied, ‘don’t worry about it, there’s nothing to know’—I had become discouraged at that time with the social sciences.

‘When we met, Robin Room asked me for a job and being an honest man said “but I know nothing about sociology”. I replied, “don’t worry about it, there’s nothing to know”—I had become discouraged at that time with the social sciences.’

A: What was ARG’s affiliation?

WC: ARG also changed its affiliation several times. It was part of the State Department of Health, at another time it became part of University of California, and then later it became part of the Medical Research Institute of San Francisco. All that was essentially paperwork, and it had little to do with any of the work that was being conducted. The affiliation was a matter of convenience and financial necessity in terms of receiving grants and the administration of grants. The center of funding of alcohol studies during the later 1970s became the National Institute on Alcohol Abuse and Alcoholism (NIAAA). ARG had other financial odds and ends, such as state funding for small projects, but almost all the funding came through the NIAAA. They were establishing other research centers all around the country, and soon the studies began to take on a world-wide status. We became familiar with people all over the world, learning who they were and about what they were doing.

A: And that search for a bimodal distribution?

WC: As the studies progressed we confirmed that we could not find a bimodal distribution for alcohol consumption. The distribution was continuous, and there was no discreet group easily defined by a pattern of use of any kind. We were also surprised with the fact that the United States was unique when compared to most other industrialized nations. In the United States there is a very substantial group of people who do not make any use of alcohol. We found that here in Berkeley, and what we found for Berkeley did not differ greatly from what was later found in studies for the nation as a whole. There were many abstainers, about 30% of US adults, which is a higher proportion than what was found in studies in Scandinavia, Canada, France, Britain and Germany. Many people reported that they did not drink because their families did not drink, because their religion forbids it, because drinking was morally wrong. Genevieve Knupfer found those people in Berkeley, which is a fairly liberal place and not far from the biggest wine-producing region in the nation.

A: The social contest for abstention?

WC: We began to learn that many people tended to associate in their lives with others much like themselves. They went to churches with people who did not drink, they went to parties, when they went at all, with people who did not drink. They did not very often have heavy drinkers among their friends and acquaintances. Reversing that coin, heavy drinkers reported that they associated with people who drank a great deal, that they went to parties and taverns, and that they did not know many people who were abstainers. This was surprising. I thought I knew my countrymen fairly well and I was wrong. Many of them were abstainers on principle and very militantly so.

A: How did this large number of abstainers and lack of a bimodal distribution affect the studies being conducted at the time?

WC: When we were asking people about drinking, most of us were worried that they would not answer questions truthfully. The years went by and the studies being conducted described the uses of alcohol and also began to ask about alcohol problems. But asking about alcohol problems and drunkenness worried us even more. We thought that people would refuse to answer the questions and perhaps break off the interviews, and the whole study would come crashing down. Interestingly enough, it was our timidity that was the problem, not the willingness of respondents to talk about all kinds of things. Were their comments on their behaviors accurate? That was a concern then as I suppose it is a concern now, and all the years in between. Some things led us to wonder. For instance, if we added up all the alcohol that people said they drank and compared it to all that was sold, there was a discrepancy with the summed interview reports accounting for only 50% of what had been sold. So, we knew that there was under-reporting.

A: And the accuracy of the reporting on problems?

WC: In theory, I suppose you could check on drink driving arrests and that sort of thing, but in practice it would be an enormous task. To make a long story short, when we did check, while some records showed events for a respondent that the respondent did not report it was also the case that respondents reported some events such as arrests or other embarrassing occurrences that the records did not show. It was probably the case that the best single source of information that you could obtain was by asking the people themselves what they did. We were, as survey research always is, dependent upon what people can and will tell you. That is not very different from the way we live our lives in general, but the worry was still with us.

A: Where did your ideas on drinking problems come from?

WC: Almost all our early questionnaires could trace their history to Jellinek’s analysis of AA members. Jellinek’s [6] book on the disease concept of alcoholism was also enormously useful because he described alcoholism, defined it, identified several types, and in so doing provided the field with a number of testable propositions. He described loss of control as marked by unintended, unwanted drunkenness. He stated that in his view this was a disease that one either had or did not have, an all-or-none type of thing. He also described craving and the fact that alcoholism was marked by this progression in phases that lead to blackouts and tremors, and that this was common among these people who were marching in the inexorable step-by-step progression into alcoholism: a disease.

A: What did you find?

WC: In the general population we found to our surprise that loss of control, blackouts and so on were not to be found only in middle-aged people. These symptoms were most common among young men. We also found that it was not a one-way type of process. These symptoms came and went; so loss of control might have been reported by someone at one time but not at a later follow up. The notion of a single progression, of a clear-cut disease, fell apart in our hands. There were alcohol problems, and serious ones, in great abundance but no single entity that could be identified clearly as alcoholism.

‘The notion of a single progression, of a clear-cut disease, fell apart in our hands. There were alcohol problems, and serious ones, in great abundance but no single entity that could be identified clearly as alcoholism.’

A: As you speak, it seems to me that the research carried out at the beginning, the California Drinking Practice Study, was very empirical and not theoretically oriented.

WC: That is right; certainly there were alcohol problems in abundance in the general population. But these problems did not seem to fit any regular pattern and if you tried to define alcoholics or severe problem drinkers by one set of criteria you would end up with a different group than if you used a different set of criteria. Also, the rates of alcohol problems and the correlates of alcohol problems would differ greatly depending on which set of criteria you chose to use. Changes in criteria to define alcoholism can be seen clearly in the evolution of the Diagnostic and Statistical Manual (DSM) of the American Psychiatric Association. Loss of control became impairment of control, and the line dividing alcoholics and non-alcoholic is no longer a sharp one but a blurred sort of thing. You can have more of it or less of it. You can be in full control or you are unable to control or you can be anything in between. Jellinek proposed a very reasonable sort of thing, but others proposed reasonable but different things. Also, craving was important in the early studies but could not be measured objectively.


  1. Top of page
  5. References

A: How did these two ‘surprises’, that is, the high proportion of abstainers in the United States and the distribution of alcohol problems in the general population, influence the overall direction of research in general population then?

WC: To make sense of what happened in those years, do you recall that there is a book by Cahalan [7] and another by Cahalan & Room [8], that focused not on alcoholism as a disease but on problems of drinkers and problem drinking? These two books were reflections of what was found in general populations: many alcohol problems, no clear-cut identifiable entity of pathology, but just a great deal of drinking that led to serious consequences. Some of these consequences were physical, many of them emotional or legal or financial. We began to think about under what circumstances different consequences would appear, which eventually led me and some others to look at the social circumstances of drinking. That is, what are the social forces that bring about this distribution of alcohol use, with its problems and with its benefits as well? This goes back to the mid-1950s and the work of Ledermann [9], in France. As you know, he then made the outrageous proposition that if one knew the per capita alcohol intake for a population one could predict the proportion of alcoholics, or of cirrhosis deaths in a population. To everybody’s surprise, it emerged that there was some relationship. There is a predictable relationship between per capita consumption and the rates of cirrhosis and all kinds of severe alcohol problems. The question then arises, why should it be true? After all, we can imagine a population in which there are many heavy drinkers and many light drinkers and a certain mean for that group. Or we can imagine a different population having exactly the same mean, but with no heavy drinkers and no light drinkers but everybody in the middle. That is not what is found in practice. Empirically, what is found is a distribution that rises rapidly towards the mean of consumption and then a tailing-off to the right, and in the tail of that curve consumption is higher and the problems are more abundant. Many people were wondering about the explanation for the regularity that seemed to underlie the distribution seen across different populations. We thought that there had to be something in the social circumstances of alcohol use that explained the mean consumption and the proportion of people in a population falling on either side of that mean, as well as the proportion at the tail end of the curve that contains most of the alcohol problems. So we began to be concerned with circumstances of use, with the norms governing these circumstances of alcohol use. How much was customary to drink at a party if you are a Californian, or a New Yorker, or a Norwegian or whatever.

‘There is a predictable relationship between per capita consumption and the rates of cirrhosis and all kinds of severe alcohol problems.’

A: So the origin of those questions on social norms and how much is customary to drink by different people is a result of the preoccupation with the social circumstances of drinking?

WC: Let me put it this way. First of all, as everybody will recognize, there are social circumstances in which one should not be found drinking or perhaps not drinking very much, and there are some where it is perfectly normal to drink; and at least for some people, to become inebriated is perfectly acceptable. College drinking parties are one such setting. But in almost any tavern can be found, at least on Friday and Saturday nights, a number of people who will be fairly well under the influence of alcohol and not view their behavior as in any way reprehensible. There is a range of expectations of how much one can acceptably use and how much is too much. These social circumstances that regulate drinking behavior can be described and are remarkably consistent in the population, becoming a kind of a social lid that keeps behavior within acceptable boundaries, with perhaps occasional lapses. That varies from situations that are very wet, in our terms, to ones that are very dry. We studied everything from spending a quiet evening at home, to how much is normal to drink when friends come over to visit, or how much is customary at parties or when you are going to drive a vehicle. In general, women are not supposed to be as drunk as men. Alcohol is more approved for younger people, and less approved for older folks. Certainly, religious groups have their own feelings about this sort of thing. I, and many others, were arguing that these were the social ties that govern our alcohol use, and ultimately produced what Ledermann had noted.

A: When you look back to the past almost 50 years, what do you see that is different in the alcohol research field now?

WC: Well, let us begin with the funding. Funding was a very difficult proposition in the 1960s, not only for us but for anyone interested in alcohol studies, because these studies were expensive. Also, we were competing with all kinds of other health-related projects funded by the National Institute of Health or the National Institute of Mental Health for the same research dollars. The notion that you were going to have any funding on any continuing basis was not at all certain. No one counted on it. We thought in terms of a year’s study, a 2-year study. This all changed little by little, but especially with the creation of the NIAAA, which led to funding on the level that permitted research centers, and people like, us to exist and young people to enter this field with the reasonable expectation of spending their working lives in it in a successful way. That was not the case when I began. Something of the same kind led to increased and sustained funding in industrialized countries all around the world at about the same time.

A: The birth of the alcohol researcher as a specialist?

WC: In the early years we were statisticians or we were anthropologists, or we were sociologists or we were psychiatrists. The notion that we were alcohol researchers had not dawned on anyone. There was no one such as an alcohol researcher. By 1970, certainly there were people who defined themselves in that way and who were thought of as alcohol researchers as though it was a specialty occupation, a way of life. That is clearly the case now.

A: In a way, the field became much more structured?

WC: It became structured. It developed a culture of its own. It became inhabited by the great figures of the past and the contemporary ones. It is a separate discipline, I think, quite separate from, for instance, psychology or sociology or what you will. Certainly, it is a specialty in the public health field.

A: Now let me gain your perspective on an area that has changed quite fast recently: the present ability to conduct quantitative analysis.

WC: I did not have a calculator for the first 5 years. Not only did I not have one, nor did anyone else. We used slide rules, and all the information was contained on huge stacks of cards which were very carefully preserved in a box in a back room. Eventually we obtained data tapes. The notion of being able to perform multivariate analysis, even though the techniques were available, the time and money to do them simply made them impossible.

A: Do you think that all these changes are for the better?

WC: I think that one of the biggest changes for the better is this: that there are many publications, good publications in which a reasonably uninformed reader can become very well informed by simply picking up various journals, reading carefully about all the debates that are presented there, the arguments, how things are going on. When I began this was not the case. There were few good journals. Most of them were not very good and to gain an education was something you pretty much had to do on your own.

A: Other changes?

WC: Another change is in the nature of some debates in the field. Jellinek’s concept of the disease of alcoholism no longer looks like a very useful theoretical scheme. Today, there are no people, as far as I know, who are arguing as they were in the early years, that a Freudian explanation of alcohol abuse was central to understanding this problem. There are no arguments, so far as I am aware, from the behavioral people arguing that alcoholism is simply a conditioned response, and all we have to do is make these guys sick every time they see a drink and they will go on and do something else. And there is no argument, as was the case with the AA people in particular, and the medical people who shared their point of view, that this is something you are born with, and that becoming an alcoholic is almost inevitable; something like an inherited ability to play the piano or read music and all you had to do was be introduced to the piano, do a little bit of practice and you would be doing wonderfully well as a musician or as a drunk because it was in your bones. All these notions seem to me to have become attenuated.

A: If you look back now, what would you say have been the contributions from survey research?

WC: I think almost everyone involved in alcohol studies would agree that we now have a fairly accurate description of alcohol use in almost all the industrialized nations in the world. All this, plus knowledge as to the consequences of various patterns of alcohol use, was lacking 25 years ago. Where it will go from here I have no idea. I think the basic sciences will play a greater role than they did but they, too, have undergone an enormous transition in the period of time and have found things that were hidden and were only speculated about. Alcohol researchers now know a great deal more about the behavior of people that was simply speculated about in years gone by. The descriptions, based on general population research, make it harder to hold to a dogmatic position. The abundance of information that the world is awash with, things that seem to be constant, that seem to make sense, seem to persist over a time-frame very, very well—that has changed the scene.

‘The descriptions, based on general population research, make it harder to hold to a dogmatic position.’

A: Is there anyone that you think influenced your work more than anybody else during your career?

WC: No one person. There were many people who were important, but it was a network of people pretty much all over the country and later all over the world who were conducting the same kinds of things. There were many publications from ARG and others which led to the exchange of ideas back and forth: for instance, the publishing of ARG’s The Drinking and Drug Practices Surveyor. There was collegiality around that document and others like it in which we influenced each other. We shared a common task in a sense, and we all had a great deal in common.


  1. Top of page
  5. References
  • 1
    Riley J., Marden C. The social pattern of alcoholic drinking. Q J Studies Alcohol 1947; 8:26573.
  • 2
    Riley J., Marden C., Lifshitz M. The motivational pattern of drinking. Q J Studies Alcohol 1948; 9:35362.
  • 3
    Maxwell M. Drinking behavior in the State of Washington. Q J Studies Alcohol 1952; 13:21939.
  • 4
    Straus R., Bacon S. Drinking in college. New Haven: Yale University Press; 1953.
  • 5
    Mulford H., Miller D. (1964) Drinking and deviant drinking, U.S.A. Q J Studies Alcohol 1963; 25:534650.
  • 6
    Jellinek E. M. The disease concept of alcoholism: New Brunswick, NJ: Hillhouse Press; 1960.
  • 7
    Cahalan D. Problem drinkers: a national survey. San Francisco, CA: Jossey Bass; 1970.
  • 8
    Cahalan D., Room R. Problem drinking among American men. Monograph no. 7. New Brunswick, NJ: Rutgers Center of Alcohol Studies; 1974.
  • 9
    Ledermann S. Alcool, alcoolisme, alcoolisation, vol. 1. Paris: Presses Universitaires de France; 1956.